WMOD00423

• Title: Currents…Developing Technology for Tomorrow's Challenges
• Date: 1/1/1993
• Weather Modification - Doc Type: Report

Subsurface models are inherently am-
biguous. They represent interpretations of
geological data that are projected down
from the surface, geophysical data that
have large uncertainties, and borehole data
that must be interpolated between widely
spaced points. Investigators often experi-
ence pressure to produce one "best-guess"
model, when a variety of models actually
fit the available data. Combining statistical
interpretations of the different data sets
would reduce this ambiguity. Many subsur-
face interpretations can exist for one set of
data (figs. 1 and 2). The number of possi-
ble subsurface configurations decreases
when two data sets are collected because

Feet

I I
a). 0 100 250
("-' ( :r.'

. 1 t \1\\ ~Tl"J( \GJ-' H'K.~

...
. G.,

I. -,., II

G.,

... (:..~

e'l' .1.

.. ,I....' i

.. (.: ..-,.

c.' ".'

" "'.I~

5

only the interpretations that fit both data
sets are possible subsurface configurations.
Investigators can further reduce the num-
ber of possible subsurface configurations
by adding descriptions of the type of geo-
logic environment, such as stream of lake
deposited soils. Data integration reduces
the ambiguity of the subsurface interpreta-
tion. Each addition of another data set to
the site information base decreases the
number of possible subsurface configura-
tions (fig. 3).
The amount of data needed to suffi-
ciently characterize ground water or
contaminated sites is not clear. The new
software, UNCERT, will develop tech-
niques for (1) integrating available data at
a contaminated site and (2) automating the

b).

I.:r~'

. (".I~ .

. If.\I'-\ '.;rOR~(iJ I"....

~
... t :.,:-
(.:.-'

(;''::UI

C.'C'" i
.

. t:t.!'

".' ,...

C; I' .,:

1,1 III"

. 1.5.1-

Figure 1. When describing the subsurface, bore.hole data are often used to interpret site conditions.
The data, however, may not be conclusive. In this case, there are 11 data samples; 6 clay, and 5 gravel.
As can be seen, although data for each map are identical, the nature of the geology in each map is sub-
stantially different. This illustrates that given data from the subsurface, there is uncertainty associated
with the interpretation of the character of the subsurface in locations that have not been sampled.

a).
( ':"
. II
.
....
..

.-:....
...

\... \ ~.I'fUl \l.t "1''\ "h."

" , I..,. ~ (";.::~
'" 11.11111 P!1m

. I;: iI:'l:

(I~

I

...

(ira,,'1 Ifill ppnl

10 ppm

( ;/'11\1'1

~ Gravel /
(;ra\cl .. ~ ~
.... ....................~ ~'..
.--:::.:....-.'.."'" ..,- ::3,
Contaminant Plume Map . . 1 ppm .
..,..;.......;.;.;.;...;.....;..........;...........;.....;.......... "

.. (:Ill)

b).

( 1.1:.

..(:.~'"

II \h\l<1'RU(.\(;1 I\'h'
. ('I
..... . ,.

... ("'..:
1111'11111

C.i ..- 1-.

1111111 PI"'t
." "( 'Ja..'
t,rau" IlHI n.." n I
r". PIIi'1

..
, ,1,1"'1

. ..

1.11'1'"1

....

(;rll\"t'1

r ,1,,\11

.......---..- .
i I .:= uninant Plume ~
1..-... ~ .......... ..

I '11111

... (:1_,

Figure 2. Based on the geologic descriptions mapped in figure 1, contaminants migrating through each
setting would behave differently. When designing a remediation plan, it is Important to know which sce.
nario will occur.